6 files changed, 88 insertions, 72 deletions

diff --git a/fs/notify/inotify/inotify_fsnotify.c b/fs/notify/inotify/inotify_fsnotify.c
index 47cd258fd24d..5dcbafe72d71 100644
--- a/fs/notify/inotify/inotify_fsnotify.c
+++ b/fs/notify/inotify/inotify_fsnotify.c
@@ -62,13 +62,14 @@ static int inotify_handle_event(struct fsnotify_group *group, struct fsnotify_ev
         event_priv->wd = wd;
 
         ret = fsnotify_add_notify_event(group, event, fsn_event_priv);
-        /* EEXIST is not an error */
-        if (ret == -EEXIST)
-                ret = 0;
-
-        /* did event_priv get attached? */
-        if (list_empty(&fsn_event_priv->event_list))
+        if (ret) {
                 inotify_free_event_priv(fsn_event_priv);
+                /* EEXIST says we tail matched, EOVERFLOW isn't something
+                 * to report up the stack. */
+                if ((ret == -EEXIST) ||
+                    (ret == -EOVERFLOW))
+                        ret = 0;
+        }
 
         /*
          * If we hold the entry until after the event is on the queue
diff --git a/fs/notify/inotify/inotify_user.c b/fs/notify/inotify/inotify_user.c
index f30d9bbc2e1b..dc32ed8323ba 100644
--- a/fs/notify/inotify/inotify_user.c
+++ b/fs/notify/inotify/inotify_user.c
@@ -386,6 +386,7 @@ void inotify_ignored_and_remove_idr(struct fsnotify_mark_entry *entry,
         struct fsnotify_event *ignored_event;
         struct inotify_event_private_data *event_priv;
         struct fsnotify_event_private_data *fsn_event_priv;
+        int ret;
 
         ignored_event = fsnotify_create_event(NULL, FS_IN_IGNORED, NULL,
                                               FSNOTIFY_EVENT_NONE, NULL, 0,
@@ -404,10 +405,8 @@ void inotify_ignored_and_remove_idr(struct fsnotify_mark_entry *entry,
         fsn_event_priv->group = group;
         event_priv->wd = ientry->wd;
 
-        fsnotify_add_notify_event(group, ignored_event, fsn_event_priv);
-
-        /* did the private data get added? */
-        if (list_empty(&fsn_event_priv->event_list))
+        ret = fsnotify_add_notify_event(group, ignored_event, fsn_event_priv);
+        if (ret)
                 inotify_free_event_priv(fsn_event_priv);
 
 skip_send_ignore:
@@ -568,7 +567,7 @@ static struct fsnotify_group *inotify_new_group(struct user_struct *user, unsign
 
         spin_lock_init(&group->inotify_data.idr_lock);
         idr_init(&group->inotify_data.idr);
-        group->inotify_data.last_wd = 0;
+        group->inotify_data.last_wd = 1;
         group->inotify_data.user = user;
         group->inotify_data.fa = NULL;
 
diff --git a/fs/notify/notification.c b/fs/notify/notification.c
index 521368574e97..3816d5750dd5 100644
--- a/fs/notify/notification.c
+++ b/fs/notify/notification.c
@@ -153,6 +153,10 @@ static bool event_compare(struct fsnotify_event *old, struct fsnotify_event *new
                                 return true;
                         break;
                 case (FSNOTIFY_EVENT_NONE):
+                        if (old->mask & FS_Q_OVERFLOW)
+                                return true;
+                        else if (old->mask & FS_IN_IGNORED)
+                                return false;
                         return false;
                 };
         }
@@ -171,9 +175,7 @@ int fsnotify_add_notify_event(struct fsnotify_group *group, struct fsnotify_even
         struct list_head *list = &group->notification_list;
         struct fsnotify_event_holder *last_holder;
         struct fsnotify_event *last_event;
-
-        /* easy to tell if priv was attached to the event */
-        INIT_LIST_HEAD(&priv->event_list);
+        int ret = 0;
 
         /*
          * There is one fsnotify_event_holder embedded inside each fsnotify_event.
@@ -194,6 +196,7 @@ alloc_holder:
 
         if (group->q_len >= group->max_events) {
                 event = &q_overflow_event;
+                ret = -EOVERFLOW;
                 /* sorry, no private data on the overflow event */
                 priv = NULL;
         }
@@ -235,7 +238,7 @@ alloc_holder:
         mutex_unlock(&group->notification_mutex);
 
         wake_up(&group->notification_waitq);
-        return 0;
+        return ret;
 }
 
 /*
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
index b619d6b8ca43..98ef624d9baf 100644
--- a/fs/xfs/linux-2.6/xfs_sync.c
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -708,6 +708,16 @@ xfs_reclaim_inode(
         return 0;
 }
 
+void
+__xfs_inode_set_reclaim_tag(
+        struct xfs_perag        *pag,
+        struct xfs_inode        *ip)
+{
+        radix_tree_tag_set(&pag->pag_ici_root,
+                           XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino),
+                           XFS_ICI_RECLAIM_TAG);
+}
+
 /*
  * We set the inode flag atomically with the radix tree tag.
  * Once we get tag lookups on the radix tree, this inode flag
@@ -722,8 +732,7 @@ xfs_inode_set_reclaim_tag(
 
         read_lock(&pag->pag_ici_lock);
         spin_lock(&ip->i_flags_lock);
-        radix_tree_tag_set(&pag->pag_ici_root,
-                        XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
+        __xfs_inode_set_reclaim_tag(pag, ip);
         __xfs_iflags_set(ip, XFS_IRECLAIMABLE);
         spin_unlock(&ip->i_flags_lock);
         read_unlock(&pag->pag_ici_lock);
diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h
index 2a10301c99c7..59120602588a 100644
--- a/fs/xfs/linux-2.6/xfs_sync.h
+++ b/fs/xfs/linux-2.6/xfs_sync.h
@@ -48,6 +48,7 @@ int xfs_reclaim_inode(struct xfs_inode *ip, int locked, int sync_mode);
 int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
 
 void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
+void __xfs_inode_set_reclaim_tag(struct xfs_perag *pag, struct xfs_inode *ip);
 void xfs_inode_clear_reclaim_tag(struct xfs_inode *ip);
 void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, struct xfs_perag *pag,
                                 struct xfs_inode *ip);
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index 34ec86923f7e..ecbf8b4d2e2e 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -191,80 +191,82 @@ xfs_iget_cache_hit(
         int                     flags,
         int                     lock_flags) __releases(pag->pag_ici_lock)
 {
+        struct inode            *inode = VFS_I(ip);
         struct xfs_mount        *mp = ip->i_mount;
-        int                     error = EAGAIN;
+        int                     error;
+
+        spin_lock(&ip->i_flags_lock);
 
         /*
-         * If INEW is set this inode is being set up
-         * If IRECLAIM is set this inode is being torn down
-         * Pause and try again.
+         * If we are racing with another cache hit that is currently
+         * instantiating this inode or currently recycling it out of
+         * reclaimabe state, wait for the initialisation to complete
+         * before continuing.
+         *
+         * XXX(hch): eventually we should do something equivalent to
+         *           wait_on_inode to wait for these flags to be cleared
+         *           instead of polling for it.
          */
-        if (xfs_iflags_test(ip, (XFS_INEW|XFS_IRECLAIM))) {
+        if (ip->i_flags & (XFS_INEW|XFS_IRECLAIM)) {
                 XFS_STATS_INC(xs_ig_frecycle);
+                error = EAGAIN;
                 goto out_error;
         }
 
-        /* If IRECLAIMABLE is set, we've torn down the vfs inode part */
-        if (xfs_iflags_test(ip, XFS_IRECLAIMABLE)) {
-
-                /*
-                 * If lookup is racing with unlink, then we should return an
-                 * error immediately so we don't remove it from the reclaim
-                 * list and potentially leak the inode.
-                 */
-                if ((ip->i_d.di_mode == 0) && !(flags & XFS_IGET_CREATE)) {
-                        error = ENOENT;
-                        goto out_error;
-                }
+        /*
+         * If lookup is racing with unlink return an error immediately.
+         */
+        if (ip->i_d.di_mode == 0 && !(flags & XFS_IGET_CREATE)) {
+                error = ENOENT;
+                goto out_error;
+        }
 
+        /*
+         * If IRECLAIMABLE is set, we've torn down the VFS inode already.
+         * Need to carefully get it back into useable state.
+         */
+        if (ip->i_flags & XFS_IRECLAIMABLE) {
                 xfs_itrace_exit_tag(ip, "xfs_iget.alloc");
 
                 /*
-                 * We need to re-initialise the VFS inode as it has been
-                 * 'freed' by the VFS. Do this here so we can deal with
-                 * errors cleanly, then tag it so it can be set up correctly
-                 * later.
+                 * We need to set XFS_INEW atomically with clearing the
+                 * reclaimable tag so that we do have an indicator of the
+                 * inode still being initialized.
                  */
-                if (inode_init_always(mp->m_super, VFS_I(ip))) {
-                        error = ENOMEM;
-                        goto out_error;
-                }
+                ip->i_flags |= XFS_INEW;
+                ip->i_flags &= ~XFS_IRECLAIMABLE;
+                __xfs_inode_clear_reclaim_tag(mp, pag, ip);
 
-                /*
-                 * We must set the XFS_INEW flag before clearing the
-                 * XFS_IRECLAIMABLE flag so that if a racing lookup does
-                 * not find the XFS_IRECLAIMABLE above but has the igrab()
-                 * below succeed we can safely check XFS_INEW to detect
-                 * that this inode is still being initialised.
-                 */
-                xfs_iflags_set(ip, XFS_INEW);
-                xfs_iflags_clear(ip, XFS_IRECLAIMABLE);
+                spin_unlock(&ip->i_flags_lock);
+                read_unlock(&pag->pag_ici_lock);
 
-                /* clear the radix tree reclaim flag as well. */
-                __xfs_inode_clear_reclaim_tag(mp, pag, ip);
-        } else if (!igrab(VFS_I(ip))) {
+                error = -inode_init_always(mp->m_super, inode);
+                if (error) {
+                        /*
+                         * Re-initializing the inode failed, and we are in deep
+                         * trouble.  Try to re-add it to the reclaim list.
+                         */
+                        read_lock(&pag->pag_ici_lock);
+                        spin_lock(&ip->i_flags_lock);
+
+                        ip->i_flags &= ~XFS_INEW;
+                        ip->i_flags |= XFS_IRECLAIMABLE;
+                        __xfs_inode_set_reclaim_tag(pag, ip);
+                        goto out_error;
+                }
+                inode->i_state = I_LOCK|I_NEW;
+        } else {
                 /* If the VFS inode is being torn down, pause and try again. */
-                XFS_STATS_INC(xs_ig_frecycle);
-                goto out_error;
-        } else if (xfs_iflags_test(ip, XFS_INEW)) {
-                /*
-                 * We are racing with another cache hit that is
-                 * currently recycling this inode out of the XFS_IRECLAIMABLE
-                 * state. Wait for the initialisation to complete before
-                 * continuing.
-                 */
-                wait_on_inode(VFS_I(ip));
-        }
+                if (!igrab(inode)) {
+                        error = EAGAIN;
+                        goto out_error;
+                }
 
-        if (ip->i_d.di_mode == 0 && !(flags & XFS_IGET_CREATE)) {
-                error = ENOENT;
-                iput(VFS_I(ip));
-                goto out_error;
+                /* We've got a live one. */
+                spin_unlock(&ip->i_flags_lock);
+                read_unlock(&pag->pag_ici_lock);
         }
 
-        /* We've got a live one. */
-        read_unlock(&pag->pag_ici_lock);
-
         if (lock_flags != 0)
                 xfs_ilock(ip, lock_flags);
 
@@ -274,6 +276,7 @@ xfs_iget_cache_hit(
         return 0;
 
 out_error:
+        spin_unlock(&ip->i_flags_lock);
         read_unlock(&pag->pag_ici_lock);
         return error;
 }